Using Genomics and Biomarkers to Predict Response

Using Genomics and Biomarkers to Predict Response

The Clinical Science Symposium “Immunotherapy: Now We’re Getting Personal—Using Genomics and Biomarkers to Predict Response” on June 5 highlighted the ongoing search for biomarkers that can predict response to immunotherapies, with research analyzing the potential biomarker value of mismatch repair (MMR)-deficient colorectal cancer, mutational load in metastatic urothelial carcinoma, and next-generation sequencing in melanoma.

Treatment with pembrolizumab, an anti–PD-1 immune checkpoint inhibitor, was found to be highly active in MMR-deficient colorectal cancer, according to Luis A. Diaz, Jr., MD, of Johns Hopkins University, who presented updated results (Abstract 103) and expanded data on a phase II trial that was published last year.1

The trial had two cohorts. Cohort A included 28 patients with MMR-deficient colorectal cancer, which is associated with hereditary nonpolyposis colorectal carcinoma. Cohort B included 25 patients with MMR-proficient disease. All patients received 10 mg/kg of pembrolizumab every 14 days. Patients had previously been treated with two or more therapies and had a performance status of 0 or 1.

Dr. Diaz said that tumors that have genetic defects in MMR pathways harbor hundreds to thousands of somatic mutations, particularly in the regions of repetitive DNA, or microsatellites. The abundance of mutant-associated neoantigens results in the tumor appearing foreign to the host immune system. “[This] foreignness results in an inflamed microenvironment with a very high expression of immune checkpoints. Treatment of these tumors with anti–PD-1 unlocks a potential antitumor response in microsatellite instability tumors,” Dr. Diaz said.

The trial endpoints were response and progression-free survival (PFS) at 20 weeks; secondary endpoints included disease control rate, PFS, overall survival (OS), and safety. Median time to follow-up was 8.7 months.

Treatment with the single agent resulted in an objective response rate of 57% and a disease control rate of 89% among the 28 patients with MMR-deficient disease; among the 25 patients with MMR-proficient disease, the objective response rate was 0% and the disease control rate was 16%. PFS and OS have not yet been reached in the MMR-deficient group. For the MMR-proficient group, PFS was 2.3 months and OS was 5.98 months. Currently, five patients, or 18% of the MMR-deficient cohort, have reached the 2-year mark, and anti–PD-1 has been held, with these patients continuing to be under active surveillance.

Future directions include completing confirmatory registration studies in first- and third-line colorectal cancer, and exploring the possibility of histology-independent indications for treatment of minimal residual disease tumors with PD-1 blockade, Dr. Diaz said. The molecular etiology of primary and secondary resistance in tumors with minimal residual disease that have been treated with PD-1 blockade also should be investigated. Questions remain about how to treat the patients that reach the 2-year mark on treatment with PD-1 blockade, including whether therapy should continue or be discontinued.

Urothelial Carcinoma

Dr. Jonathan E. Rosenberg
Jonathan E. Rosenberg, MD, of Memorial Sloan Kettering Cancer Center, presented results of a trial that examined immune and genetic predictors of response to atezolizumab in metastatic urothelial carcinoma (mUC; Abstract 104). Atezolizumab was the first U.S. Food and Drug Administration–approved PD-L1 inhibitor and previous trials demonstrated its efficacy in mUC.

The exploratory analyses were performed on archival tumor specimens from patients with mUC. Predictors of response to atezolizumab were identified as PD-L1 immune cell (IC) status (p = 0.0109); The Cancer Genome Atlas (TCGA) subtype (p = 0.0384); and mutational load (p ˂ 0.0001). PD-L1 IC combined with TCGA subtype significantly improved on PD-L1 IC alone or subtype alone, and the three-biomarker combination of PD-L IC plus TCGA subtype plus mutational load significantly improved on the combination of PD-L1 IC plus subtype (Fig. 1).

“Simultaneous assessment of these characteristics may define drivers of immune responsiveness to inform potential combination of strategies,” Dr. Rosenberg said. “These data highlight the importance of the interaction between the tumor and its microenvironment in understanding response to atezolizumab.” Patients with mUC who are treated with atezolizumab “can experience durable response across all investigated biomarker subgroups including PD-L1 IC status. The IMvigor201 exploratory biomarker studies suggest that atezolizumab efficacy is driven by genomic, molecular, and immunologic factors related to adaptive immunity.”


Discussant Alexandra Snyder Charen, MD, of Memorial Sloan Kettering Cancer Center, noted that the primary message of these two studies is that “patients whose tumors harbor the highest mutational burdens exhibit a substantial improvement in OS. Conversely, those patients not in the highest mutational burden group do not fare as well and, thus, demonstrate the subgroup that we need to focus on for ongoing clinical studies.”

She pointed to several unanswered questions, including why an elevated mutation burden matters. The current hypothesis is that mutations generate relevant neoantigens, she said. Neoantigens, or neoepitopes, occur when mutations are transcribed and translated into abnormal proteins, which are then processed and thought to be presented by the tumor or the antigen-presenting cell to the immune system.

“It is not entirely clear whether an elevated mutational burden results in a large neoantigen burden or increases the chances of forming a few critical neoantigens,” Dr. Charen said. Data from other labs have suggested that the “few critical neoantigens” scenario is more likely, and some data suggest that the clonal neoantigens are the relevant ones. “In addition, the relative contributions of MHC Class II and tumor-associated antigens in this scenario have not been fully elucidated, but studies are ongoing,” she said.

A better understanding of the factors that lead to primary and acquired resistance is also needed. “How do we consider tumor heterogeneity and integrate data from the biologic complexity at the tumor site? How do we boil down this complexity into usable biomarkers?” Dr. Charen said.

PD-1 Expression in Bladder Cancer

Douglas Buckner Johnson, MD, MSCI, of Vanderbilt University Medical Center, presented results of a study that used hybrid capture–based next-generation sequencing (HC NGS) in melanoma to identify markers of response to anti–PD-1/PD-L1 (Abstract 105). The results indicated that mutational load as determined by this platform “may aid in treatment stratification,” he said.

Anti–PD-1 therapy produces a response in about 30% to 45% of patients with advanced melanoma, but biomarkers of response are needed. The study’s aim was to profile a smaller fraction of the genome that could potentially serve as an accurate surrogate for mutational load. Using HC NGS, the trial sought to assess whether the number and/or type of mutations correlates with outcomes to agents blocking PD-1 (aPD1) in melanoma, he said.

The sequencing on 236 to 315 genes was performed on archival samples from patients who had been treated with aPD1 to correlate mutational load and specific mutations with clinical outcomes and to compare mutational load as calculated by whole-exome sequencing. The correlation of T-cell receptor sequencing was also examined, Dr. Johnson said.

The initial cohort included 32 patients and the validation cohort included 33 patients. In the initial cohort, patients who had a response to aPD-1 had a higher mutational load compared with those who did not have a response (median, 45.6 vs. 3.9 mutations/megabases; p = 0.003). In the validation cohort, the median mutational load for responsive disease was also higher (37.1) than for nonresponsive disease (12.8; p = 0.002). Relative risk, PFS, and OS were superior in patients with high mutational load compared with patients with an intermediate/low mutational load (p ˂ 0.001). Responders also had more cytosine-to-thymine transitions (median, 33.5 vs. 3.0).

Melanomas with NF1 mutations had a high mutational load, whereas BRAF-mutant, MRAS-mutant, and BRAF/NRAS/NF1 wild-type melanomas had lower median mutational load, Dr. Johnson said. Neither T-cell receptor clonality nor PD-L1 gene amplification correlated with response in the archival samples.

PD-1/PD-L1: Predictive Biomarkers?

Discussant Padmanee Sharma, MD, PhD, of The University of Texas MD Anderson Cancer Center, said that expression of a single biomarker for selecting patients for treatment “may not be feasible,” given that immune responses are dynamic and evolve over time.

Studies up to this point “have identified potential prognostic biomarkers, including tumor-infiltrating T cells, immune gene signatures, PD-L1 expression, TcR repertoire, and mutational load, which indicate patients that are likely to have a better response to therapy. A single predictive biomarker has not been identified that can be used to select and/or exclude patients for immune checkpoint therapy, so we have to be very careful how we look at these individually,” Dr. Sharma said.

“We were very much benefited by the era of genomic medicine, where we all looked at a single mutation and chose a drug based on that single mutation. Immune responses are dynamic and we are unlikely to have a single biomarker. Because immune responses are evolving and dynamic over time, expression of a single biomarker to select patients for treatment will really not be feasible and we should not continue on the single biomarker path,” she said.

“Both PD-L1–negative and PD-L1–positive tumors can respond to treatment because PD-L1 expression is dynamic. Biomarker development for immune checkpoint agents will require integration of multiple biologic components, such as CD8-infiltrating T cells, PD-L1 expression, IFNγ-signature genes, mutational load, etc., as opposed to a single molecule,” she said.

Importantly, Dr. Sharma said, “we need to consider longitudinal evaluation of tumor tissues throughout treatment to define the adaptive pathways that will need to be targeted to elicit durable responses. Currently, we are using a lot of archival samples that may not give us an accurate picture of what is occurring within the tumor microenvironment.”

–Kathy Holliman, MEd